The exploratory observational study was conducted between April and June 2023 after receiving a positive ethics approval from the Ethics Committee of Heidelberg University Hospital (S-233/2023, April 26, 2023), using a document analysis (32) and SWOT analysis in an online group discussion as a business method for identifying a company’s strategic need for action (33). Its use is also well established in medicine (34–36). The observation period within the document analysis and SWOT analysis performed covers the start dates of the networks (RN May 2021; LU January 2022) until the implementation of the study, the start of which was defined by the presence of a positive ethics vote in April 2023.

For the document analysis, the documents created during the establishment and realization of the network activities were retrospectively evaluated. The documents were publicly available on the respective websites. Other documents, such as minutes and internal progress reports, were analyzed retrospectively in an anonymized form after receiving the ethics vote.

The target group of the SWOT analysis were the network partners of RN and LU, who offer a health and/or care service in the ambulatory sector, or patients who are involved in the network advisory board/in the network organization. The target group of both networks (n = 23 Rhein-Neckar; n = 52 Ludwigsburg) was contacted by e-mail and invited to participate in the SWOT analysis. A reminder was sent after a few days. In case of interest, an information leaflet and a consent form were sent. Information about the study procedure was provided verbally and open questions could be asked and clarified. In case of consent, socio-demographic data (gender, role in the network, rural/urban work location) were pseudonymized. Consent to participate was given in writing. In case of more than eight interested persons per network, a purposive sampling strategy was planned. As this number was not reached in either network, all interested participants were invited to participate. An incentive of € 150 was paid for participation in the SWOT analysis in an online group discussion (90 min).

In the case of RN, in the document analysis an internal progress report, programs of the training courses, protocols of the advisory board meetings and working groups, and contents of the website were evaluated (29). In LU, relevant documentation from the Rhein-Neckar internal progress report, the protocol of the introduction of the SWOT group discussion, the internal documentation from LU, and website content (30) were included in the analysis. In the document analysis, using a combination of qualitative content analysis and thematic analysis with support of MAXQDA, topics for the presentation of the networks were identified by SS in an iterative process and discussed with LG (master’s student of health services research and implementation science, experience in qualitative research). The contents were integrated by SS (general practitioner, experienced qualitative researcher of the study team and coordinator of RN) supported by the main coordinator of LU (JK) based on the results of the document analysis in an iterative process. The identified topics and contents were additionally checked for plausibility in the network coordination (UM, gastroenterologist, experienced researcher of the study team and coordinator of RN) and for comprehensibility in the study team (LG, SV).

The SWOT analysis was moderated in an online group discussion by a member of the study team and recorded pseudonymously by another person. Due to the coordinating role of SS in RN, she did not participate in the corresponding regional group discussion in order to avoid social desirability. The group discussion was structured as follows: 1. exchange in small groups as introduction; 2. input to the SWOT analysis; 3. SWOT analysis in small groups with documentation in a 4-field board; 4. presentation of the SWOT analysis results of the groups; 5. common discussion of the synopsis of the analysis results as well as completion of the documentation if necessary. From the documentation of the group discussions (SWOT documentation and protocol of the discussion), topics were 1. merged and 2. clustered by SS. Subsequently, the result was checked for comprehensibility by LG and aspects with difficulties in understanding and/or comprehensibility were discussed point by point and agreed between SS and LG. Finally, the synthesized document was sent to the group of participants in the form of a member check. The original SWOT documentation was added. As reflection questions were asked: “1. In your opinion, has something important been lost? If so, please name the aspect(s). 2. Is there anything mentioned in the synthesized version that you see as a wrong result of the group discussion? If so, please identify this aspect/these aspects. Please only comment on what was documented in the group discussion and do not add any new aspects.” The feedback was discussed, agreed upon and integrated between LG and SS.

Identified categories from the document analysis were start, sponsoring, build-up, coordination, consented goals, definition, activity status, activities, and treated patients.

Table 1 shows the characteristics of the two networks RN and LU. RN started on May 22, 2021, LU started on January 01, 2022. There were similarities and differences, but the latter predominated: Similarities can be seen in the definition of the network. Both networks defined network participants who have agreed on common goals as well as patient pathways and have established ways of communicating, organizing and collaborating. On the one hand, identified network partners in both networks are listed on a website and are thus contact persons for medical care in Long COVID; on the other hand, general practitioners in both networks are responsible for basic care and care coordination. In RN, patient representation was continuously involved in the coordination of the network through the advisory board via self-help group members; in LU, the network was defined by stakeholders. In addition, perceived care needs were the trigger for the establishment of the networks in both regions. There were also similarities in the agreed objectives, such as avoiding underuse and overuse. The regional collaboration included working groups, advisory board meetings, participation of regional stakeholders in training programs (RN) and quality circles, interface agreements within the care process, and participation of regional actors in training programs (LU). At the time of manuscript preparation in May 2023, there were approximately the same number of listed network partners with multidisciplinary and multispecialty composition, i.e., 34 (RN) or 36 (LU) stakeholders who are actively involved in the care of patients with long COVID. Differences can be seen in the start of network activity, which started 7 months earlier in RN than in LU. RN received project funding that included network coordination and reimbursement for SWOT analysis, whereas LU was carried out solely on a voluntary basis. In both networks, care was provided within the standard of care without additional incentive. The perceived need that triggered the establishment of the network in RN was from an inpatient perspective (that means special ambulatory department for Long COVID of the university hospital) and in LU from an ambulatory perspective, which continued in the further establishment and coordination. In the consensual objectives, RN showed a more provider-oriented perspective, whereas in LU the patient level was also taken into account. In the category definition of the network, the unique selling point in RN was the formulated conditions of participation and the presentation of an interdisciplinary, interprofessional and intersectoral advisory board with integration of the meso level; in LU, more concrete and specific coordinated agreements for the interfaces emerged. The activity status in May 2023 shows differences between the regions. The project character in RN includes on the one hand an end of the project and on the other hand the prospect of a continuation in a follow-up project. In LU there is a continuing activity without funding. Differences are still evident in the intersectoral focus in RN and the ambulatory focus in LU.

The network activities are presented in Table 2, which shows many similarities but also differences: In particular, in RN the focus is on the involvement of the specialized sector of the university hospitals, on continuing education and on the supraregional approach, whereas in LU the regional, ambulatory, concrete level of care is visible through the interface agreements and the derivable patients treated in the network. Topics from the stakeholders’ perspective about the patients treated in the network were derived from the advisory board protocols. They describe in terms of severity a wide range from mildly affected with temporary reduction in performance to severely affected with long term suffering. Topics within the described group of severely affected include post-exertional malaise, use of non-established therapies, such as apheresis, and resulting social difficulties up to unemployment due to sickness. Furthermore, a sense of desperation and lack of care was perceived by some stakeholders. The stakeholders reported a high time demands for care. A repeatedly discussed topic were aspects of the psyche in the disease pattern Long-COVID, including the themes patients with and without previous mental illness, overlap with psychosomatic illnesses, and patient concerns about psychologizing. Challenges in attributing the patient-presented symptoms were also repeatedly addressed.

A total of n = 14 participants (n = 7 RN, n = 7 LU) took part in two regionally separated online group discussions. There were n = 6 male participants. A more urban place of work was indicated by n = 8 and a more rural place of work by n = 6. Participants included seven physicians (one outpatient rehabilitation physician, four general practitioners, one pediatrician, one rheumatologist), three occupational therapists, one speech therapist, one physiotherapist, one psychoneurologist and one psychotherapist.

The strengths and weaknesses identified by the participants are presented in Table 3. In both networks, the multidisciplinary-multispecialty character is seen as a strength, in RN also the intersectoral approach. Both networks valued the resulting collaboration. In terms of activities and effects on medical care, differences between the network functions became clear, with RN emphasizing the strength in the area of knowledge transfer and LU emphasizing the strength in the area of concrete contact persons, increasing of caregivers’ motivation, improved interface exchange and more quickly appointments. Regarding the weaknesses mentioned, there was a high level of agreement and consensus about the lack of reimbursement for participation and the limited participation. There was also agreement on the high efforts of care and inadequate reimbursement, as well as the perception that there were still gaps in knowledge and motivation among colleagues. Differences arose in the assessment of collaboration, with RNs reporting a lack of concrete action and realization, which was not an issue in LU.

The opportunities and risks identified by the participants are presented in Table 4, which shows a high degree of agreement in the thematic areas. Both networks continued to see a need for intervention in the area of long COVID, with increase of public awareness and mobilization of external funding opportunities seen as relevant. The groups also agreed on the potential transferability of the network approach to other diseases, e.g., Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS), and in LU the promotion of engagement with a “difficult topic” that is often met with rejection was emphasized. In RN, the role of the university hospital in science and networking was seen as an important opportunity, and in LU it was pointed out that experience in the ambulatory sector could support research. Further development opportunities, such as intersectoral care concepts and institutionalization, were also pointed out. Risks identified by both networks ranged from limited resources, such as high effort for integrative care activities to the risk of project termination without sufficient support and recognition. There was a consensus that the network activity could lead to other diseases receiving too little attention. The establishment of alternative care structures was viewed differently, with RNs seeing this as a potential risk and LUs seeing the potential inhibition of such structures as a potential risk. Different lenses were also used to illuminate the issue of mismanagement of care, with attention drawn in RN to the relevance of long COVID subgroups and in LU to the potential mismanagement of patients with other health problems by such a network offering.

A strength of the study is the written presentation of identified themes and strategic need for action of the realized networks as ad hoc best practice examples despite limited resources in terms of time and content. The pragmatic and quick approach of document analysis and SWOT analysis with an explorative character offers the possibility to quickly generate initial strategic hypotheses on the topic of long COVID networks.

Document analysis has advantages such as being an efficient method, availability and cost effectiveness, and limitations such as lack of detail, low retrievability and biased selectivity (32). The potential bias of using the people involved in the coordination was countered by the exchange with LG as an independent member of the study team.

The SWOT analysis is a method for strategy development, but it does not replace established social network analysis or quantitative or qualitative methods at the patient, professional (inpatient and outpatient) and system levels, which should follow in the next step. Due to the study design, the present work does not claim to be representative. The number of participants (n = 14) is limited, but at the same time the statements included a broad spectrum, as urban and rural workplaces, men and women, different professional groups and two different regions in southern Germany were surveyed. A selection bias may have occurred due to the participation of particularly motivated or positively minded individuals. The invited patient representatives from the advisory board of the RN network could not participate due to time or health reasons. One-sided data collection from stakeholders can lead to bias, so the results must be clearly interpreted as a survey from their perspective. Including patients’ experiences is important (8), and should be considered in future studies.

A SWOT analysis is used to identify strategic needs for action (30). The results of this process in the described regional networks showed on the one hand indications that network participation could contribute to a rapid learning and resilient health system coping with long COVID, for example through the reported resulting collaboration, contribution to knowledge transfer, and improvement of care for individual patients. On the other hand, at the same time, barriers such as lack of reimbursement, high efforts of care, and persistent motivation gaps became apparent. Potentials mentioned were, e.g., transferability to other diseases such as ME/CFS, promotion of addressing a “difficult topic” and promotion of intersectoral care concepts; risks mentioned were, e.g., limited network resources and negative effects on the development of other structures. Resulting implications for practice and research address a call to policy makers and funders to support further research to find out what generalizable results regarding usefulness, effectiveness, and efficiency including transferability to other post-infectious diseases can be derived, what aspects best contribute to impact, what is needed for the sustainable establishment, and, in summary, generate more robust evidence. Because they are different, the pros and cons of both networks need to be considered. The application of participatory approaches involving patients and stakeholders seems reasonable and timely (55).